[Project Page] [Challenge Page] [Demo Program] [Presentation]
- Subject: Free Topic (Food-organism utilization throughout the AI-based aquaculture industry)
- Home page: https://sarc.jnu.ac.kr/contest/20211105/
- Motivation:
- Stable mass feeding management of food organisms is importance of 'artificial seed culture' industry
- Problems cause serious economic loss in the aquaculture artificial seed production industry
- Reduction of aquaculture food organisms (plankton)
- Difficulty in managing mass culture/feeding of food organisms
- Mass mortality in the process of seed production
- Decreased utilization of food organisms by field
A sharp decline in marine food organisms (plankton) due to global warming and marine environmental pollution
- It is necessary to select and concentrate the government R&D AI data and technology to solve fundamental problems such as instability and low productivity in the aquaculture industry
- Team Name: ADLER
- Affiliation: Chonnam National University, South Korea
- Please see in our project page at [Project Page]
project
├── aquaculture
│ ├── app_v2
│ ├── apps
│ ├── assets
│ │ ├── cache
│ │ ├── data <-- Setup Data step
│ │ │ └── final_info.csv
│ │ └── models
│ ├── exps
│ ├── utils
│ ├── cli_main.py
│ └── common.py
│ └── ...
├── data
│ ├── a2i_data <-- copy csv, 먹이생물 into here
│ │ ├── csv
│ │ │ ├── 10월01일
│ │ │ │ ├── 2-1-1-1-1-1001-0010000.csv (id-code.csv - sensors data)
│ │ │ │ └── 2-1-1-1-1-1001-0020000.csv
│ │ │ ├── 10월04일
│ │ │ └── ...
│ │ └── 먹이생물
│ │ ├── 10월01일
│ │ │ ├── 고성
│ │ │ │ ├── 2-1-1-2-2-1001-0120001.jpg (id-code.jpg - microsopy images)
│ │ │ │ ├── 2-1-1-2-2-1001-0120002.jpg
│ │ │ │ └── ...
│ │ │ ├── 일해
│ │ │ │ ├── 2-1-1-2-2-1001-0110001.jpg
│ │ │ │ ├── 2-1-1-2-2-1001-0110002.jpg
│ │ │ │ └── ...
│ │ ├── 10월04일
│ │ └── ...
│ ├── preprocessed <-- Setup Data step
│ │ ├── full_info.hdf5
│ │ ├── full_info.xlsx
│ │ ├── final_info.csv
│ │ ├── final_info.xlsx
│ │ └── final_info.hdf5
│ └── exps
└── images
- Install Anaconda3 at https://www.anaconda.com/products/individual
- Activate environment base
# Linux
conda activate base
# Window
activate base
- Create environment a2i
conda create -n a2i python=3.8
- Activate environment a2i
# Linux
conda activate a2i
# Window
activate a2i
- Install requirements packages at environment base
pip install -r requirements.txt
- Copy csv (sensors data), 먹이생물 (microscopy images) to folder data
- Open console
- Go to project root
# Linux
cd <project dir>
# Window
cd /d <project dir>
- Activate Environment a2i
# Linux
conda activate a2i
# Window
activate a2i
- Generate index files
python aquaculture/cli_main.py index
python aquaculture/cli_main.py detect-all
- Go to project root
- Activate Environment a2i
- Type command
python aquaculture/cli_main.py app2 --app-type dash
- Open Web browser and type url: https://localhost:8050
- Go to project root
- Activate Environment a2i
- Type commands
- Generate index files
python aquaculture/cli_main.py index
- Detect number of cells in a microscopy image
python aquaculture/cli_main.py detect-one --id-code <id_code> (2-1-1-2-2-1001-0120126)
- Detect number of cells in all microscopy image and save to index file
python aquaculture/cli_main.py detect-all
- Data analysis belongs to places, grouping by day
python aquaculture/cli_main.py data-analysis
- Training and Evaluating baseline algorithms
python aquaculture/cli_main.py baseline --config <config file>
- Training and Evaluating tabnet algorithms
python aquaculture/cli_main.py tabnet --config <config file>
- Prediction food-organism quality from sensor data
python aquaculture/cli_main.py prediction
--model <model_name> (sklearn, tabnet)
--model-path <the path of model weights>
--id-code <id_code>
or
python aquaculture/cli_main.py prediction
--model <model_name> (sklearn, tabnet)
--model-path <the path of model weights>
--temp <temparature>
--do <dissolved oxygen>
--ph <pH>
--sal <salinity>
--ntu <nephelometric turbidity unit>
- Open html files in notebooks folder to view results of console tasks
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
All material is made available under Creative Commons BY-NC-SA 4.0 license by Adobe Inc. You can use, redistribute, and adapt the material for non-commercial purposes, as long as you give appropriate credit by citing our paper and indicating any changes that you've made.